Artificial intelligence in hybrid imaging: Enabling or restricting the workforce?

The nuclear medicine environment is now undoubtedly considered to be a truly digital imaging community.  A community whereby technology exists and practitioners are integral to ensuring image optimization / acquisition, accurate processing / quantification and in the case of hybrid imaging, using the dual modality approach to further enhance the patient's diagnosis.

The pace of introducing intuitive software platforms within nuclear medicine is something which needs to be carefully considered.  This is especially the case for hybrid imaging practice, where a 'push and pull of data' culture has begun to evolve.  The hybrid imaging practitioner faces a number of challenges in terms of ensuring their role definition, development and deployment is recognized and central to the patient's journey.  Creating an environment that embraces technology, but does not allow it to control the role is a crucial element of future workforce training and there needs to be an understanding that multiple pressures exist and may impact on the creation of 'domain ownership' by the nuclear medicine workforce.  The figure below outlines the necessary pathway that needs to be undertaken by practitioners as they emerge as a new sub-society within the nuclear medicine fraternity. 

Copyright Marc Griffiths 2013



My previous blogs have explored the notion of creating a clear pathway of training and education for the nuclear medicine practitioner working within a hybrid imaging environment.  Unless there is a clearly defined role, working in a patient centric approach, technology will dominate the delivery of services.  The use of systems workflows and seamless integration of various aspects of the nuclear medicine examination can be viewed as enabling the workforce, however there is also limited potential for 'flat collaboration' as previously discussed by Fridell (2009) and cultural lag, as practitioners fail to understand how technology can be embraced and used to enhance service delivery. 

Knowledge rather than reliance on policies

Despite having an apparent abundance of policies and procedures at their fingertips and visible within the clinical imaging environment, nuclear medicine practitioners often appeared to ask their colleagues for advice or feedback.  This was most noticeable in situations where experienced nuclear medicine practitioners would be approached by staff who were not as familiar with hybrid imaging techniques or technical aspects (e.g. Quality Control checks) and were asking for help / guidance.  Although there was access to a range of policies and other useful documents, the practitioners generally preferred to ask their colleagues and learn from the experience.  This often involved citing previous experiences and / or bringing up example studies to demonstrate particular aspects of image acquisition or processing etc. 

The documents pertaining to hybrid imaging practice were being produced by a mixture of professional disciplines, including hybrid imaging practitioners and clinical scientists.  There also appeared to be a notion of health competition between the professional disciplines in terms of who could produce the most documentation / policies in some incidences.  Why was this the case? This behaviour appeared very 'tribal' in nature, almost mirroring graffiti street tags and highlighting a claim on the technology that was being used. 

Issues however may arise when protocols change or new imaging pathways are introduced.  How is this information disseminated through the hybrid imaging community? This also leads to a potential lack of evidence based knowledge development and practitioners not being able to develop enquiry based learning approaches within the workplace.

SUV calculations and decision making abilities

From my time in clinical PET, I've noticed a variation in the calculations of Standard Update Values (SUV) between different manufacturers software platforms. I'm not sure why this occurs and perhaps it could be down to the use of different calculation algorithms being utilised.

Practitioners, physicists and consultants need to be aware however of the potential for variations in SUV calculations between different manufacturers. It would be interesting to conduct a validated comparison project between different vendors software to assess the potential variations in calculations.  This also highlights to me the need of ensuring practitioners have a strong understanding of quantification within nuclear medicine and in particular hybrid practice.  Ensuring the skills, training and education frameworks are in place is integral to the future of effective / innovative service provision.